I'm running MonteCarlo simulations to compute the most likely voltage values in a grid. The process is to generate a random load from a load values distribution, and simulate the load with the power flow simulation. Sometimes for some loads combination, the power flow does not converge (but still provides a value that, far from perfect is still usable)
When the MonteCarlo run goes well I get this:
However when a MonteCarlo run contains evaluations that did not converge I get this:
I want to understand this phenomenon better, so any insight or some reference to read further is welcome.